Re-evaluating currently available data and suggestions for planning randomised controlled studies regarding the use of hydroxyethyl starch in critically ill patients - a multidisciplinary statement

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Introduction: Hydroxyethyl starch (HES) is a commonly used colloid in critically ill patients. However, its safety has been questioned in recent studies and meta-analyses. Methods: We re-evaluated prospective randomised controlled trials (RCT) from four meta-analyses published in 2013 that compared the effect of HES with crystalloids in critically ill patients, focusing on the adherence to 'presumably correct indication'. Regarding the definition of 'presumably correct indication', studies were checked for the following six criteria (maximum six points): short time interval from shock to randomisation (<6 h), restricted use for initial volume resuscitation, use of any consistent algorithm for haemodynamic stabilisation, reproducible indicators of hypovolaemia, maximum dose of HES, and exclusion of patients with pre-existing renal failure or renal replacement therapy. Results: Duration of fluid administration ranged from 90 min up to a maximum of 90 days. Four studies considered follow-up until 90-day mortality, three studies 28-/30-day mortality, whereas four studies reported only early mortality. Included studies showed a large heterogeneity of the indication score ranging between 1 and 4 points with a median (25%; 75% quartile) of 4 (2; 4). Conclusions: The most important question, whether or not HES may be harmful when it is limited to immediate haemodynamic stabilisation, cannot be answered yet in the absence of any study sufficiently addressing this question. In order to overcome the limitations of most of the previous studies, we now suggest an algorithm emphasising the strict indication of HES. Additionally, we give a list of suggestions that should be adequately considered in any prospective RCT in the field of acute volume resuscitation in critically ill patients.

This publication has 37 references indexed in Scilit: